Brave's AI assistant now integrates with PDFs and Google Drive

Original Article ↗ Hacker News Discussion ↗

Brave’s new Leo assistant, which can summarize and interact with PDFs, Google Drive files and web pages, is prompting mixed reactions about the future of AI in browsers. Many commenters appreciate features like built-in ad blocking and potential productivity gains, but others raise concerns about privacy, data being sent to remote LLMs, Brave’s past crypto- and advertising-related controversies, and the broader trend of VC-backed Chromium forks that may later “enshittify.” Firefox frequently emerges as the preferred alternative for those worried about Chromium dominance and tighter integration of AI and tracking into core browsing.

Browser preferences and alternatives

  • Many users report moving away from Chrome, especially due to YouTube ads and perceived site breakage, even with uBlock Origin.
  • Firefox is frequently recommended as the main non‑Chromium alternative: praised for longevity, privacy stance, extensions (especially uBlock Origin), and as a counterweight to Chromium’s dominance.
  • Criticisms of Firefox include weaker PWA support, Android performance and bugs, awkward tab UX without extensions, and search-engine shortcut ergonomics vs Chromium.
  • Brave gets mixed reviews:
    • Positives: fast, strong built‑in ad/tracker blocking, Tor integration, Web3/IPFS options, fairly seamless YouTube experience, decent sync, vertical tabs.
    • Negatives: crypto/BAT model seen by some as a grift, past missteps (affiliate code insertion, tipping UX), AI/crypto “bloat,” and general distrust of a VC‑backed for‑profit.
  • Other mentioned options: Edge (good features, lower memory than Chrome), Cromite/Ungoogled Chromium (privacy-oriented, community builds), LibreWolf (hardened Firefox), Safari (likened to “new IE” by some; others use it happily).

AI assistants in browsers

  • Some see AI-in-browser as redundant “chatbot everywhere” marketing, driven more by investor optics than user need.
  • Others find clear value:
    • Summarizing articles and long videos.
    • Comparing products or real-estate listings.
    • Organizing tabs and aiding memory over previous browsing.
    • Potential for advanced content filtering (e.g., filtering out certain article types or images).
  • Concerns:
    • Sends page and document contents to remote LLMs, undermining Brave’s “security/privacy first” image.
    • Fear that such integrations normalize surveillance and could be used to bypass ad blockers or deepen DRM/lock‑in.
    • Some users simply find current AI unreliable (e.g., failing simple spreadsheet queries).

Privacy, monetization, and trust

  • Debate over how Brave funds these features:
    • Some point to BAT/crypto and its ad network as inherently suspect but note that user rewards are small and features are mostly optional/opt‑in.
    • Others note Brave can reduce LLM costs via open‑source models on own hardware and intends paid tiers with usage limits.
  • Strong skepticism that user data won’t be used for training, despite claims; some “guarantee” the opposite.
  • Several commenters express that, if AI is to be integrated, they would trust Mozilla/Firefox more than Brave or other Chromium forks.

Desire for better history/memory tools

  • Multiple users want AI or extensions that index local browsing history, bookmarks, and PDFs to help rediscover past information, with preferences for:
    • Fully local or self‑hosted processing.
    • Fine‑grained control (e.g., only bookmarked pages).
  • Some small existing tools and projects are mentioned, but this is seen as an underdeveloped area compared to generic page-summarizing chatbots.